The Cost of Doing Business

A couple of weeks back, I was on the receiving end of my company’s first summons. It was a trivial issue that resolved itself within a couple of days. But it gave me some insight into how my customers (typically, large companies) think about these things.

My first reaction was shock (“How could this happen? There must be some mistake”). That feeling was soon eclipsed by outrage (“This is ridiculous, we haven’t done anything wrong”); which was followed by regret (“I wish we had just avoided this situation”); finishing up with irritation (“I can’t believe I have to waste time on this when I have so much real work to do.”)

When I mentioned this reaction to a couple of our customers, they just chortled to themselves and suggested that I get used to it: as your business grows, they said, you can be certain that more of these will follow.

That’s when it struck me: dealing with these issues – and by implication, e-discovery – is by no means unusual; it has become part of the cost of doing business. In the same way that companies pay their taxes or process employee visas, they respond to subpoenas, demand letters, and regulatory inquiries. Whether they themselves are directly implicated in wrongdoing, or they were innocent bystanders who had nothing to do with it, doesn’t make any difference. They have to do the work all the same.

With this in mind, I feel better prepared for the next summons, whenever it comes. Right now, we are focused on recruiting and training; at some point, if all goes well, we will get to e-discovery.

Open Platforms in E-Discovery

Most large companies face a dilemma. Should they open up their products and invite others to build features on them, creating a “platform” or ecosystem around themselves? Or would that be inviting the proverbial fox into the hen-house, meaning they should instead prevent others from integrating with their product or leveraging it to create add-on functionality?

In the internet world, there is no doubt about the answer: throw open the doors via easy-to-use APIs (“application programming interfaces”) and let a thousand flowers bloom. That’s what FaceBook did a couple of weeks back with their announcement of the FaceBook Platform, and it has already led to hundreds of new applications for their users. It is what Skype did so effectively, creating a mini-industry around themselves of voicemail, skins, ring-tones, and more. Even eBay, which has jealously guarded its feedback ratings and has habitually crushed smaller companies in its cross-hairs, is embracing the open platform mantra, announcing this week that third-party companies can build features that work with eBay in new ways.

By contrast, telecom companies live in a world of closed standards. Even in the wireless industry, which is arguably the most competitive part of the telecom world, the carriers (Cingular, T-Mobile, Verizon, etc.) exact a heavy toll on any application trying to reach their handsets. As friends in the industry tell me, “There’s a reason why there has never been a billion dollar mobile application company.”

In e-discovery, the large technology vendors like EMC, HP, Symantec, and ZANTAZ face the same choice. Their email archiving products store huge amounts of data. Should they let 3^rd party e-discovery software analyze that data, giving their customers more choice? Or should they slam the door shut, and try to force customers to use their own proprietary e-discovery applications?

The answer, it seems, depends on what they want to be when they grow up. As the market leader, Symantec is confident enough to open its archive (Enterprise Vault) to 3^rd party applications while offering customers its own Discovery Accelerator for litigations holds and some document review. Similarly, perhaps because of its powerful brand, HP focuses on storage optimization with HP RISS and partners with e-discovery software, often with huge savings for its customers. On the other side of the coin, smaller companies like ZANTAZ and Mimosa see themselves as e-discovery companies: they seek to leverage their storage products to get customers to also buy their e-discovery applications.

In the long-run, my feeling is that any archive of any stature will have to adopt open standards. Customers will demand it, and (unlike telecom companies) the archive vendors do not have the market power to resist. Over time, they will also come to appreciate (as HP and Symantec do now) that enabling 3^rd party applications to analyze the data they store is to their advantage, since it creates a powerful, additional incentive to store more information in the archive.

What Web 2.0 Applications Can Teach Enterprise Software

The other day, I came across the fascinating statistic that over 50% of products returned every year to stores across America have absolutely nothing wrong with them. Apparently, consumers used them for an average of 20 minutes and then gave up, because they were too complicated.

At this point, most customers of traditional enterprise software could be forgiven for thinking: “I wish I could do that.” Enterprise applications are notoriously feature-laden, complicated to use, and difficult to install. They make their users feel stupid, by presenting them with complex pictures that look like amoeba or toolbars with 150 different options. Why does enterprise software seek to punish its customers in this way?

Partly, because customers ask for it. Whether they are buying a dishwasher or an accounting application, people habitually over-estimate their ability to figure out how a complicated product works and, as a result, pay more for features that they never use. Partly, it’s because enterprise software is designed by engineers who think everyone is as technically proficient as they are, and by marketing people who view every additional feature as a new selling point.

By contrast, Web 2.0 applications such as FaceBook, Flickr, StumbleUpon, or Meebo are incredibly easy use. Even an idiot who has never seen these applications before can use them without an instruction manual or a training course. You could say that’s because they are trivially simple applications. But I think it’s primarily because, if they were not so easy to use, people would simply click away and try something else – i.e., they would die.

That to me is the real lesson that Web 2.0 apps can teach enterprise software: make something that is easy to use, easy for someone to install, and easy for them to evaluate. Get people addicted to your application because it’s so good (the average FaceBook user spends 4+ hours a day on the site). No doubt, this is harder to do with enterprise applications because they are inherently more complex. But figure out a way to hide the complexity, packaging all the functionality users need into a design that’s easy to use. This is a key characteristic of e-discovery software applications; it's the genius of salesforce.com's CRM application and Apple's iPod; and, it needs to be a core skill of any company creating enterprise applications today.

What is E-Discovery 2.0?

In a previous post, I wrote about the forces transforming e-discovery, a phenomenon that has received increasing attention from the press, most recently in this week’s Economist magazine. While everyone agrees that something big has changed, and (generally speaking) on the reasons why, people struggle to put their finger on exactly what e-discovery has become.

That’s why I think the concept of “E-Discovery 2.0” is so helpful. Analogous to Web 2.0, E-Discovery 2.0 is a set of new processes, technologies, and services that enable companies to manage huge volumes of data, lower costs, and meet tight deadlines.

New Processes

When e-discovery meant handing over a few boxes of paper, companies did not need much of a process. But in today’s world, where it involves terabytes of data, teams of reviewers, and precious little time, it is a very different story. To cope with the growing volume and complexity of e-discovery issues, companies have had no choice but to adopt new processes. These include:

• Collect and Preserve: Most companies have now established procedures so that, when the need arises, they can collect all data relevant to a case and ensure that it cannot be changed or deleted.

• Analyze Up Front: When presented with more work than can be done, a company’s only option is to work smarter, not harder. That means analyzing the collected data up front, to cull it down to only those emails and documents directly relevant to the case at hand.

• Collaborate Efficiently: E-Discovery has become a team sport. And whenever you have a team, you need a playbook, or a process, to ensure work is not repeated and that everyone is marching towards the same goal.

New Technologies

If technology created this problem, by making electronic communication so pervasive and voluminous, then it can also solve it. In recent years, several new technologies have arisen that enable companies to store and sift through their data to fulfill e-discovery obligations. The most significant of these trends include:

• From tape to disk: As the cost of disk storage has continued to decline, more and more companies are abandoning tapes and instead keeping their data online. Email archiving software optimizes for storage efficiency, allowing companies to keep hundreds of terabytes of data readily available for e-discovery.

• From search to analysis: Basic keyword search has evolved into sophisticated analysis technology that mines email meta-data for relevance, links messages together into discussion threads, and groups them by topics. These analysis applications allow users to sift through millions of messages in minutes, to rapidly identify, tag, and export relevant data.

• From closed systems to open standards: Until recently, technology providers made no effort to integrate their applications, leaving customers to fend for themselves. But that has started to change. Symantec Enterprise Vault and HP RISS now have open APIs, creating pressure on others to follow suit. George Socha’s Electronic Discovery Reference Model (EDRM), a standards body, has received widespread support, accelerating progress towards creation of an open e-discovery platform.

To anyone working in litigation support, legal, or information security, all this is quite unremarkable. Of course they use technology to address e-discovery. Obviously, there has to be a process. From the company’s perspective, e-discovery has become no different to HR or finance – it is a core competency, part of doing business.

And that, perhaps, is the most remarkable thing about E-Discovery 2.0 – in only a few short years, it has become so widespread and deeply entrenched within the enterprise, that people barely notice it.

Can E-Discovery Really Be That Expensive?

I tend to have a "Mark Twain perspective" on statistics and apply a healthy grain of salt to any numbers quoted by analysts and industry experts. But when end-users speak, I sit up and listen. That's why I was very interested to read here that Microsoft "spends an average of US$ 20 million for e-discovery per litigation, according to one company exec." (My thanks to George for alterting me to the article)

If true, it is an astounding number - but one that is quite consistent with what we have seen first hand working with other large enterprises ourselves. Once you factor in processing costs (an average of $1,800 per GB), review costs ($200/hour), and the huge volume of information being generated and stored, you can get up to $20 million on a single case surprisingly fast.

The White House And The Problem of A Billion Emails

The other day, Michael Clark of EDDix sent me a fascinating academic paper (thanks, Michael!) about “information inflation” at its impact on the legal system. I had never really thought of it this way, but there have really only been 3 significant events in the evolution of information:

1. Writing (c. 5,000 years ago): Pre-historic man started to etch his markings on clay tablets, stone, wax, papyrus, bark, cloth, wood, paper, cave walls and anything else that came to hand.
2. Printing (c. 1450): Gutenberg’s movable type printing press enabled mass production of information, contributing to (among other things) the Renaissance and the Scientific Revolution.
3. Digitization (c. late 20th Century): The personal computer, wide area networks, internet, email, have all led to a massive explosion of information in the past 50 years. As the article points out, “close to 100 billion emails are sent daily…In a small business, whereas formerly there was usually 1 four-drawer file cabinet full of paper records, now there is the equivalent of 2,000 four-drawer file cabinets full of such records, all contained in a cubic foot or so in the form of electronically stored information.”
   

How can the legal profession cope, given that a lawyer’s job is often to synthesize this mind-boggling amount of data? Fortunately, the authors have a solution:

“A family of computer technology employing new types of search methods and techniques beyond use of mere keywords should now be considered for use in litigation….Litigators can no longer depend on manual review alone. It is too time-consuming and expensive – with cost often exceeding the amounts in dispute.”

To illustrate its point, the paper tells the story of the White House and the problem of a billion emails. During the Clinton administration, the White House agreed to a form of electronic record keeping called ARMS (Automated Records Management System). At the end of each administration, these records are handed over to the National Archives and Records Administration (NARA). The table below shows the number of stored emails NARA has, or expects to receive at the end of each administration.



Now assume that, like previous administrations, the Next President’s administration is subject to a lawsuit that requires e-discovery. The paper calculates:

“Without employing any automated computer process to generate potentially responsive documents, the review effort for this litigation would take 100 people, working 10 hours a day, 7 days a week, 52 weeks a year, over 54 years to complete. And the cost of such a review, at an assumed billing rate of $100/hour, would be $2 billion. Even, however, if present day search methods are used to initially reduce the email universe to 1% of its size (i.e., 10 million documents out of 1 billion), the case would still cost $20 million for a first pass review conducted by 100 people over 28 weeks, without accounting for any additional privilege review.”

This is a great example of why companies and government agencies are adopting e-discovery 2.0 technologies that go far beyond keyword search. In the face of information inflation, what choice do they have?

From Web 2.0 To E-Discovery 2.0

If there’s one idea that has captivated Silicon Valley in the past 3 years, it is Web 2.0. People may debate its meaning and definition, but the gist of it is clear: a handful of powerful forces have coalesced to make the internet of today fundamentally different to what it was 5 years ago. Opinions vary on which of these forces is most important: the growth of broadband to the home; open source, ajax and other technologies which lower the cost and increase the functionality of web applications; the power of community in a world where more people are on the web. Whichever you choose, there is no doubt that collectively these forces have had a huge impact, powering the growth of now-household names such as Google, MySpace, and YouTube.

I believe that an analogous set of changes is transforming the way companies do e-discovery. Ten years ago, e-discovery was an after-thought – a necessary, but incidental, part of corporate legal expenses. Today, it is a huge line-item in the legal budget, a headache for corporate IT, and the foundation upon which many cases are built.

E-discovery 1.0 was an ad hoc activity; e-discovery 2.0 is a core business process. E-discovery 1.0 was barely noticed; e-discovery 2.0 is driving the news cycle, affecting everyone from Intel to the US Attorney General. In the legal world, e-discovery 2.0 has had every bit as big an impact on enterprises as Web 2.0 has had on the dating lives of teenagers.

What happened? A series of fundamental changes have made e-discovery far more important, expensive, and complex than it was in the 1990s. Chief among these changes are:

1. Email, Not Voicemail: In the past 10 years, companies have switched from voicemail to email as the primary way they communicate. This has created a written record where none previously existed. Just as oral histories eventually die out, every voicemail eventually gets deleted; but emails and the written word live forever. Whatsmore, the convenience and time-efficiency of email makes it addictive, with the result that every meaningful conversation is captured, time-stamped, and attached to a person’s name. Given that many legal cases turn on intent, and proving who knew what when, this makes email a virtual treasure trove for anyone building a case.

2. Electronic Files, Not Paper: Electronic files are fundamentally different to paper documents: they reproduce like rabbits and are far cheaper to store. For example, one laptop is the equivalent of 2,000 boxes of paper; one server corresponds to 8,000-40,000 boxes of paper. The number of servers and laptops holding vast quantities of email is only increasing as the cost of hard disk storage falls, down from $2.04 per GB in 2004 to $0.77 per GB in 2006. Net net: going electronic has vastly increased the amount of data that must be analyzed as part of the discovery process.

3. Sooner, Not Later: Recent changes to the FRCP guidelines have moved e-discovery up in the process, forcing companies to have an e-discovery plan within 99 days of a suit being filed. Since disputes rarely settle that quickly, that means enterprises must now incur the expense of e-discovery on every case, not just the small number that actually make it to court. The result is a massive increase in e-discovery expenses and workload.

Anecdotal evidence of e-discovery 2.0 is everywhere. A few years back, no one would have guessed that every major analyst firm would have people dedicated to tracking e-discovery. Nor would you have expected to find a litigation support manager at every major enterprise.

So what exactly is e-discovery 2.0? Well, I will talk about that in a future post.

eDiscovery In The Blogosphere

It has now been over a month since I started blogging about Email Intelligence and eDiscovery, and perhaps the most pleasant surprise has been to find that I am not alone. As the chart below shows, there has been an explosion of activity around eDiscovery in the blogosphere since the FRCP Rule changes on December 1, 2006, with the happy result that today there are several voices which are well worth listening to.

To assist you in your travels, I offer a brief (and by no means comprehensive) guide to the blogs which have caught my eye. In general, they fall into 3 categories:

1. Messaging Mavens: For an entertaining look at email in the news, I would suggest Roger Matus’ Death by Email, which has everything from videos to colorful commentary. In a similar vein, Chris Foreman’s Messaging Mogul offers an interesting perspective on relevant technologies, in a way that is refreshingly free of the usual mind-numbing marketing-speak.

2. Legal Eagles: There are many lawyers who blog, often covering arcane topics or issues particular to a specific industry. But the one general, business-oriented legal blog that I would recommend is Andrew Cohen’s blog, which makes a range of complicated legal topics accessible to the general reader.

3. Article Clippers: Finally, there are the folks who helpfully collect interesting articles from around the web into a single place, so you can get a filtered view of the latest eDiscovery stories. Foremost among these is Jeff Fehrman and Bob Krantz’s edd blog online which focuses on eDiscovery and forensics.

I do not pretend to have anything close to a complete list. So if there are others worthy of a mention, please add them as a comment so that I can update my list.

Go Ahead, Sue Me!

It is a truism to say that it is easier to dispense advice than to follow it, and with good reason. How many venture capital firms practice the financial discipline they preach to their portfolio companies? How many management consulting companies employ the innovative management theories they advocate to their clients? And how many technology companies actually leverage leading-edge technology to solve their own business problems?

The answer, at least based on my experience, is “not very many”. For example, if you look at Silicon Valley’s leading technology companies, the vast majority do not have an e-discovery solution in place. Yes, there are some exceptions but for the most part, when it comes to e-discovery, the likes of eBay, Google, Yahoo, and (until recently) Intel have preferred to muddle through with manual, error-prone, expensive processes.

The justifications are typically the same. Some technology companies argue that they don’t need a legal discovery solution because theirs is not a litigious industry; others say they delete everything off their Exchange servers within three weeks and so don’t have any email to discover; all agree that things like email and document retention policies are needlessly bureaucratic.

The danger of this “we- don’t- need- car- insurance- because- we- will- never- have- an- accident” approach has been brutally exposed in the past few weeks by the painful experience of Intel. In case you missed the press coverage: AMD sued Intel for anti-trust violations. Like any company on the receiving end of a subpoena, Intel was obliged to provide opposing counsel with all email and documents relevant to the case.

If Intel had an e-discovery solution, that would have been a straightforward process. Intel’s IT group would simply identify a group of messages by date range, person, and perhaps keyword within their larger email archive. The legal group would then use an analysis product to cull down the messages to only those relevant to the case. The whole thing would take a few days. But that’s not what happened. Since Intel did not have an e-discovery solution, the company had no simple way to preserve and analyze the relevant data. Intel’s legal department was obliged to inform over a thousand employees that they could no longer delete data at will. Somewhere along the line, the message did not get through and employees kept on deleting. As a result, Intel was forced to go back to the judge with the proverbial “the dog ate my homework” defense, while AMD cried foul.

How much this costs Intel is yet to be determined. But my guess is that they will end up spending more on lawyers to fix the mess than they would have spent on an e-discovery solution that would have avoided the problem to begin with.

While I have given up on venture capitalists and management consultants, I remain optimistic that the technology industry will practice what it preaches and leverage technology to solve its own business problems in e-discovery. As Intel discovered, it is not enough to have smart lawyers on staff. You also need to equip them with an e-discovery solution that allows them to preserve and analyze information relevant to the case.

To do otherwise is an open invitation to your competitors to sue you. Just ask Larry Ellison – or better yet, SAP.

Analyze Email First, Talk Later

More than perhaps any other type of case that companies deal with, employment disputes often boil down to “he said, she said”. Since witnesses are rarely present, both sides quickly then look to email for supporting evidence. This usually happens behind closed doors giving the innocent bystander no visibility into how the process works. That’s why the case of the Justice Department and the 8 fired US Attorneys is so interesting – it illustrates what happens every day in similar, less high-profile employment disputes.

The questions are always the same:

1. Who made the decision? Initially, the answer given was Kyle Sampson, the Attorney General’s Chief of Staff. But email told a different story, showing that the Attorney General was involved, something that Mr. Sampson later confirmed in Congressional Testimony.

2. Why were the people fired? Initially, the answer was “performance reasons”. But internal department e-mail messages show consideration was also given to the views of senators, administration policy priorities, and legislative goals.

3. Was the decision justified? Well, that’s where the evidence stops, and human judgment comes in. Supporters of the decision would say it was perfectly justified, but poorly executed; opponents would argue it is more evidence of politicizing the judiciary.

As in other employment disputes that occur outside of the public eye, email analysis takes the “he said, she said” out of the situation. If you know who made the decision and why, it becomes much easier to decide whether the action was appropriate. The current difficulties at the Justice Department stem as much from the fact that they did not analyze their emails before making public statements, as it does from what they actually did.